RE: SIgh [was: Irony heaped on irony]

-----Original Message-----
From: Tim Berners-Lee
To: James Tauber; xml-uri@w3.org
Sent: 5/20/00 7:46 PM
Subject: SIgh  [was: Irony heaped on irony]


>If we could please start by defining the identifier to be a URI and if
>we feel we need new URI spaces then defined them separately.
[snip]
>If there seriously is only a small minority on this list who can
>consider the simple basic architecturally straightforward step of making
>this a URI then we have a problem.

Tim Bray and David Megginson have presented very clear, concise statements
of the mainstream XML "expert opinion" here.
See http://lists.w3.org/Archives/Public/xml-uri/2000May/0210.html
and http://lists.w3.org/Archives/Public/xml-uri/2000May/0016.html
Neither seem to think that "we have a problem", even if "we" includes people
who use RDF and/or believe in the vision of the "semantic web"
If Mr. Berners-Lee and Mr. Connolly wish to convince more than a small
minority of their position, they need directly address these postings.  I
haven't seen that yet.

>Now we have very little
>respect being paid to RDF, and an attempt by XML folks to prohibit this
>behavior. Not just to not provide it, but to prohibit others from doing
>it.

I just don't follow this argument.  The Namespace spec just doesn't specify
that a relative URI used as a namespace name should be "absolutized" before
a character by character equality comparison.  How does that prohibit RDF
users or anyone else from using "strict" URIs as namespace names, or from
putting a useful resource at a physical location on the web corresponding to
that URI? 

There's no doubt that various specs need to have ambiguities clarified and
inconsistencies ironed out, but the only people seriously inconvenienced by
the Namespace spec as it stands are those who want to use relative URIs in
namespace names *and* have generic XML software do the grunt work of
absolutizing the URIs before comparing them.  Some application-specific code
could do this, and some spec that references XML (e.g. RDF) could require
*its* implementers to do this kind of thing, right?  

In other words, I see the situation from very much the opposite perspective
as Mr. Berners-Lee: XML and the namespace spec *allow* RDF users and
"semantic web" visionaries to do what they want to do, they just don't
*insist* that the rest of the XML world do things the RDF/semantic web way.